xclim 0.61.0

v0.61.0

on Python PyPI

Contributors to this version: Pascal Bourgault (@aulemahal), Trevor James Smith (@Zeitsperre), Hui-Min Wang (@Hem-W), Éric Dupuis (@coxipi).

Announcements

The next major release of xclim will be v1.0. This new version will have some significant breaking changes such as:

• Migration of xclim.indices module into the new xclim.compute module for more efficient indicator composition.
• xclim.indicators.generic refactoring to remove lots of redundant code in favour of more standardized and easier-to-use primitive functions.

Users should expect that existing scripts may need to be updated in order to continue operating as usual. The xclim developers may release some patch versions to address small issues before v1.0.
We suggest temporarily pinning your dependencies (xclim <1.0) if your workflows require significant effort to adapt to changes, particularly if they depend on direct calls within xclim.indices.
A migration guide will be made available within the official documentation.

For more information on what will comprise the next major release and some future goals of xclim v1.0, consult the meta-issue here: GH/2352.

New indicators and features

• xclim.indices.generic.doymin and xclim.indices.generic.doymax will now return nan if all values along the time axis are the same. They now also support all-nan arrays (PR/2314).
  • This changes the behaviour for indicators land.snw_max_doy, land.snd_max_doy, land.doy_qmin and land.doy_qmax.
• Added two zero_inflated arguments to xclim.indices.stats.standardized_index and xclim.indices.standardized_precipitation_index to control how zero-precipitation probabilities are handled. (GH/2279, PR/2280).
• xclim.indices.stats.parametric_pdf allows to compute PDF distributions with given input parameters and values (PR/2323).
• xclim.indices.standardized_precipitation_index and xclim.indices.standardized_precipitation_evapotranspiration_index now can accept genextreme and lognorm as inputs for dist. (GH/2326, PR/2327).

Internal changes

• Documentation adjustments. (PR/2325):
  • Updated the intersphinx directives to point at new xsdba sphinx documentation targets.
  • Updated the ReadTheDocs configuration to build docs with an older Python version and a newer conda version.
  • Updated the pre-commit hooks (check-jsonschema) in order to accept the latest supported conda version in ReadTheDocs config.
  • Set docs/conf.py to ignore sphinx_autodoc_typehints.guarded_import errors raised by xarray type guarding.
• Set SocketBlockedError to be a subset of the Exception class when pytest-socket is not installed. (PR/2324).
• Carbon and energy reporting via green-coding-solutions/eco-ci-energy-estimation has been configured to better aggregate results for the xclim main branch as well as for all Pull Requests. (PR/2090).
• xclim now has a set of guidance documents on the kinds of AI-assisted contributions that are considered acceptable and how they must be disclosed (AGENTS.md, AI_POLICY.md, new section in CONTRIBUTING.rst). (GH/2321, PR/2346).
• The README.rst file now shows the Ouranos logo. In rendered documentation, the logo style is dynamic to the light/dark theming. (PR/2349).

Breaking changes

• The choice of method "APP" was removed when using a genextreme distribution in standardized indices (xclim.indices.standardized_index and related functions). (GH/2326, PR/2327).

Bug fixes

• dist in xclim.indices.standardized_index can now be a scipy.stats.rv_continuous as previously announced. (GH/2326, PR/2327).
• sphinx-autodoc-typehints has been pinned due to recent build failures on ReadTheDocs. (PR/2090).